Import txt file and pick the values after the selected key words using Regular Expression

Hello everybody,

I have a large non-homogeneous text file. And I need find the key words in the text then pick the value next to the key words. Here is one part of this text file:

        sfhafjlhakjfhahfaoh(some text before)
        LAW NUMBER
        8907    0    1    0
           -1876.98     11440         1
        8    2    2    2    7    8
        LAW TYPE:
        152 0 1 
         7 8      
        163 154 155 156 
        Geomaterial_2 - solid PHASE                                
        ELASTIC CONSTITUTIVE LAW FOR SOLID ELEMENTS
        AT CONSTANT TEMPERATURE
        USE OF EFFECTIVE STRESSES. ISOL =    1
        NUMBER OF SUBINTERVALS.... NINTV=    1
        YOUNG'S MODULUS .......... =   0.200000E+08
        POISSON'S RATIO .......... =   0.300000    
        SPECIFIC MASS AS A MATERIAL LAW,
        RHO ...................... =    2670.00    
        LAW NUMBER
        288    0    1    0
             -13.45     110         1
        8    2    2    2    9    8
        LAW TYPE: 
        171 0 1 
         5 6     
        173 174 175 179 
        Geomaterial_2 - liquid PHASE                               
        WATER-AIR SEEPAGE- VAPOR -THERMAL COUPLED
        CONSTITUTIVE LAW FOR SOLID ELEMENTS
        ISOTROPIC CASE IANI =     0
        FORMULATION INDEX FOR krw IKW  =     0
        FORMULATION INDEX FOR kra IKA  =     0

I need the first values after the key words and the values in the third line after the key word 'LAW NUMBER' and 'LAW TYPE'. So in this case: two vector will be created: Lawnumber=[8907 288] and Lawtype=[152 288] and another two matrix of the third line will be [8 2 2 2 7 8; 8 2 2 2 9 8] for LAW NUMBER and [163 154 155 156; 173 174 175 179]

Mr.Oleg Komarov proposed me to use regexp. His code is very easy and powerful here is the lien: http://www.mathworks.com/matlabcentral/answers/13585-find-the-key-word-in-the-text-file-then-pick-the-value-next-to-it

here is the code:

% Import the whole file at once

fid  = fopen('test.txt','r');
text = textscan(fid,'%s','Delimiter','','endofline','');
text = text{1}{1};
fid  = fclose(fid);
% Parse with regexp

tk = regexp(text,'LAW NUMBER[\s\.=]+(\d+)|LAW TYPE[:\s]+(\d+)','tokens');
% tk = regexp(text,'LAW TYPE\s+(\d+ ){2}(?:[^\n]+\n){2}(\d+ )+','tokens'); Optional code
% textscan([tk{1}{:}],'%f') Optional code
% COnvert to double
tk = reshape(str2double([tk{:}]),2,[])

It is very powerful to get the first value after key words. But the optional code doesn't work very well. Until now I am not successful to get the third line. Is someone could improve it and help me out ?

Thank you very much.

Gringoire

Answer 1

[Oleg Komarov]

This time I slightly different solution:

fid  = fopen('test.txt','r');
text = textscan(fid,'%s','Delimiter','');
text = text{1};
fid  = fclose(fid);

%Parse LAW NUMBER

idx = find(~cellfun('isempty',strfind(text,'LAW NUMBER'))) + 1;
LW  = cellfun(@(x) textscan(x,'%f%*[^\n]'),text(idx),'un',0);
LW  = cell2mat(cat(1,LW{:}));
LWm = cellfun(@(x) textscan(x,'%f'),text(idx+2),'un',0);
LWm = cell2mat([LWm{:}]).';

% Parse LAW TYPE

idx = find(~cellfun('isempty',strfind(text,'LAW TYPE'))) + 1;
LT  = cellfun(@(x) textscan(x,'%f%*[^\n]'),text(idx),'un',0);
LT  = cell2mat(cat(1,LT{:}));
LTm = cellfun(@(x) textscan(x,'%f'),text(idx+2),'un',0);
LTm = cell2mat([LTm{:}]).';

[gringoire]

Hello again,

This time it works perfectly.

Many thanks.

have a great day.

Answer 2

[Fangjun Jiang]

For a file like this, I prefer using fgetl().

NumbCount=0;
TypeCount=0;
fid=fopen('test.txt');
fline=fgetl(fid);
while ~feof(fid)
    if strfind(fline,'LAW NUMBER')
        NumbCount=NumbCount+1;
        fline=fgetl(fid);
        Temp=sscanf(fline,'%d');
        LawNumber(NumbCount)=Temp(1);
        fline=fgetl(fid);
        fline=fgetl(fid);
        Temp=sscanf(fline,'%d');
        LawNumberMatrix(NumbCount,1:6)=Temp(1:6);
    elseif strfind(fline,'LAW TYPE');
        TypeCount=TypeCount+1;
        fline=fgetl(fid);
        Temp=sscanf(fline,'%d');
        LawType(TypeCount)=Temp(1);
        fline=fgetl(fid);
        fline=fgetl(fid);
        Temp=sscanf(fline,'%d');
        LawTypeMatrix(TypeCount,1:4)=Temp(1:4);
    end
    fline=fgetl(fid);
end
fclose(fid);

[gringoire]

Thank you very much. Your code works well. In fact, before I used fgetl too. But I don't know the function of feof.... It seems very useful. :P Have a good day.

[gringoire]

Hello, this is the comment for today's question. I don't know what's happened. Your answer is missing from the web...

In fact, I am trying to use the code which you gave me yesterday. Here it is my mistake. I copied a wrong file. the F = cell2mat(textscan(fid,priformat1)) should be F = cell2mat(textscan(st,priformat1))

I wanna loop each line to get the value. But as Mr.Oleg said, the %d cannot recognize a fixed width file...

Thanks for your help indeed.